Prognostic Model In Glioblastoma

ObjectivesIn order to predict the prognosis of patients with glioblastoma, we did univariate and multivariate analysis of clinical characteristics associated with prognosis, and we chose the significant clinical characteristics factors to propose the glioblastoma prognostic model, which can evaluate the prognosis of patients with glioblastoma. So the doctor could optimize the therapeutic regimen to improve survival quality.MethodsRetrospectively analyzed119patients with pathologically confirmed glioblastoma who underwent tumor resection at Xiangya Hospital between2007and2011. Univariate analysis factors of survival time was performed using Kaplan-Meier method. The significant factors found in univariate analysis were tested in multivariate analysis using the COX regression method, then we confirmed the significant clinical characteristics factors. Based on the regression coefficient and relative risk to establish a prognostic model, namely glioblastoma personal prognosis index (GBMPI). Add the scores of each factor, we could get the final scores. To study the survival time of patients with different scores, Log-Rank regression method inspected survival rate.Results(1) Median age was48years (from8to73years); there were77men and42women;29.4%of patients were ECOG3/4at first presentation;92.4%of patients were unilateral focus,38.6%had multifocal tumor;58.9%of tumor diameter were less than5cm;90.8%of patients were primary glioblastoma; accuracy of rapid pathological diagnosis was38.7%; Complete resection in98patients,69.7%had radiotherapy and chemotherapy treatment. The median lifetime was15.4months,1-year and2-year overall survival were60.5%and26.8%, separately.(2) We made univariate analysis using Kaplan-Meier method. The result showed age, ECOG score and clinical pathological classification influenced the prognosis, which had significant statistically differences. Multivariate analysis for the3variables was performed using the COX proportional hazards regression model the results showed the3factors above also had relationship with prognosis. We chose these factors as the key prognostic factors, and the prognostic model was GBMPI=Age+ECOG+Clinical Pathological Classification. According to the GBM personal prognosis index, we divided patients into different group, Log-Rank regression method examined the survival rate of different groups, the results showed statistically significant different, the lower the score, the better prognosis.ConclusionsAfter statistical analysis, the age, ECOG score and clinical pathological classification were the key population characteristic factors, based on that we established prognostic model. Using the GBMPI we can inspect clinical data and conclude the prognosis with different scores. Application of this model can better guide the doctor to predict the prognosis of patients with glioblastoma, and consult other factors to set individualized treatment regimen. Ultimately we could make more benefit for patients and improve the quality of life. ObjectivesTo investigate the expression of EGFR and P53and their potential significance in primary and secondary glioblastoma. Based on immunohistochemical staining, we wanted to establish a novel molecular pathological diagnosis to improve the GBM prognostic model made before.MethodsDetecting the expression of EGFR and P53by immunohist-ochemical staining in37cases of GBM patients and5cases of normal brain tissue. Statistical analysis was performed to investigate the relationship between the EGFR and P53expression and GBM classification, used the Fisher discriminant to establish a method of molecular pathological diagnosis, and finally reformed the GBM prognostic model.Results(1) EGFR highly expressed in primary GBM while P53highly in secondary GBM. EGFR(+)P53(-) highly existed in primary GBM but EGFR(-)P53(+) highly existed in secondary GBM. (2) GBM discriminant model:ZGBm=-1.282×EGFR+0.108×P53-0.162(if ZGBM less than Y, it was primary GBM, in contrast it was secondary GBM, but if ZGBM equal to Y, it was anyone. Besides, ZGBM was discrimnant value, Y was mean of center of the mass, EGFR and P53both were labeling index). Its effective rate was90%.(3) Advanced prognostic model was GBMPI=Age+ECOG+Molecular Pathological Classification.ConclusionsEGFR and P53highly expressed in primary and secondary GBM separately, EGFR(+)P53(-) highly existed in primary GBM while EGFR(-)P53(+) highly existed in secondary GBM. GBM discriminant model could effectively distinguish primary and secondary GBM. Advanced prognostic model could better guide the clinical doctor to predict the prognosis of patients with glioblastoma, and consult other factors to set individualized treatment regimen.